strange search results

[darionis]

Hello,

Could anyone, please, care to explain this:

[img]http://s27.postimg.org/dyewtujg3/Capture.png[/img]

I keep getting duplicate search results. Even for simple queries, i get 2 results and they are both the same node. I use Alfresco 5.0.d running from Maven SDK 2.1 with SOLR4.


I also noticed that couple of search results don't even contain "imp*" in description or content but they are still found!!??


Can anybody suggest how to use SOLR admin console to debug this? Is index corrupted?


Thanks!

[darionis]

Couple of observations:

1. If I delete SOLR data files, repository is reindexed and then search starts to work properly and the same query returns 2 results which is correct.


2. I would say that the duplicate results are caused by moving documents from folder to folder (it finds them in the both old and new folder) and incorrect results are because properties were updated and old value matches the search. I guess SOLR index was not updated correctly when dokument is moved or its property was updated. I tried to reproduce but I couldn't, it looks like it works ok now (after reindex), but I am worried it will come back.

[darionis]

OK, it is confirmed:
As soon as I move doc from folder to folder, I start getting duplicate results (the other day I wasnt able to reproduce because I had incorrect query)

I opened SOLR's index with Luke and found these 2 entries with the same DBID but different paths. Document is definitely moved and is not present in the first path but SOLR is not aware of that.

Is this a bug? Am I doing something wrong?

Why is this thread completely ignored? This looks rather serious - am I the only one with this problem?

[darionis]

This appears to be fixed for enterpise edition:
https://issues.alfresco.com/jira/browse/MNT-13767
https://issues.alfresco.com/jira/browse/MNT-14003

Is it going to be backported to community?

(I am tired of talking to myself)

[adamsz]

Hey,
did you found the solution for community version?

Adam